USB charging, part 1: requirements [LWN.net]

USB charging, part 1: requirements

Posted Jul 25, 2016 20:23 UTC (Mon) by eds (guest, #69511)
Parent article: USB charging, part 1: requirements

So If I connect a battery-powered device to a host computer, USB data communication is required before charging can begin.

But what if I don't trust the host computer? And use a cable or adapter that connects the power pins but not the data pins. Does that mean the host will (and should) refuse to supply sufficient charging current, because no device enumeration occurred?

USB charging, part 1: requirements

Posted Jul 25, 2016 22:43 UTC (Mon) by mathstuf (subscriber, #69389) [Link] (7 responses)

If you don't trust the host computer…don't plug a cable into it? If you want just power, look to the SyncStop[1] (formerly the "USB Condom").

[1]http://syncstop.com/

USB charging, part 1: requirements

Posted Jul 26, 2016 6:15 UTC (Tue) by eds (guest, #69511) [Link] (6 responses)

Yes, I'm specifically asking what happens when such a device is in use. SyncStop blocks the data pins, which is exactly what I described.

If no data pins means no enumeration, and no enumeration means the host will cut the current back, then it's not a very useful way to recharge.

USB charging, part 1: requirements

Posted Jul 26, 2016 10:48 UTC (Tue) by gioele (subscriber, #61675) [Link] (2 responses)

> Yes, I'm specifically asking what happens when such a device is in use. SyncStop blocks the data pins, which is exactly what I described.
>
> If no data pins means no enumeration, and no enumeration means the host will cut the current back, then it's not a very useful way to recharge.

What SyncStop/USB condom should do is to have a chip connected to the data pins that does the enumeration-dance with the host and, once it is done, it should forward the current to the gadget/phone/camera attached to it. In this way the enumeration is done, but not by the phone/camera itself.

USB charging, part 1: requirements

Posted Jul 26, 2016 14:12 UTC (Tue) by jezuch (subscriber, #52988) [Link] (1 responses)

> What SyncStop/USB condom should do is to have a chip connected to the data pins that does the enumeration-dance with the host and, once it is done, it should forward the current to the gadget/phone/camera attached to it.

How does it know that it doesn't send too much power for the device not to fry?

USB charging, part 1: requirements

Posted Jul 26, 2016 15:40 UTC (Tue) by jem (subscriber, #24231) [Link]

A power supply does not send power to a device, the device draws power from the supply. The voltage is fixed at 5 volts (newer versions can supply 20 volts, when negotiated), the current drawn is up to the device.

USB charging, part 1: requirements

Posted Jul 26, 2016 11:26 UTC (Tue) by mathstuf (subscriber, #69389) [Link] (2 responses)

Yes, it blocks the data pins to your device, but will negotiate on its behalf (and as a "proof" it doesn't put data through to your device, it doesn't hook the data lines up in that side).

USB charging, part 1: requirements

Posted Jul 29, 2016 5:38 UTC (Fri) by eds (guest, #69511) [Link] (1 responses)

I can find no evidence that SyncStop / USB condom has anything at all connected to the data lines. Everything I've seen describes it as a passthrough with data lines cut.

But perhaps I just haven't searched hard enough. Do you have a link? Or know of a competing product that does this?

USB charging, part 1: requirements

Posted Jul 29, 2016 5:41 UTC (Fri) by mathstuf (subscriber, #69389) [Link]

It has to communicate with the host for power. Otherwise it's just be a simple cable which doesn't route any of the data lines at all. It doesn't have data lines to the device side though.

USB charging, part 1: requirements

Posted Jul 26, 2016 0:56 UTC (Tue) by neilbrown (subscriber, #359) [Link] (1 responses)

> Does that mean the host will (and should) refuse to supply sufficient charging current

No, it means that the host *may* refuse to supply sufficient charging current.

There are lots of USB-powered thingies around, like fans for summer, hand-warming mittens for winter, and pretty blinking lights for the festive season. They don't enumerate the bus, but they work fine when plugged into most USB sockets. If you tried to plug 4 of them into an unpowered USB hub, I suspect they wouldn't work. Hopefully the hub would detect the excess current draw and turn off.

The key point about battery powered mobile devices is that sometimes you plug in a USB cable to charge, sometimes to transfer data, sometimes for both. The device needs to know what to do. If it just pulled all the current it could, like the warming mittens, it might cause the port to shut down and then you would never get your cat photos. The mittens have no cat photos, so causing a port shutdown isn't a concern.

USB charging, part 1: requirements

Posted Jul 26, 2016 6:19 UTC (Tue) by eds (guest, #69511) [Link]

Thanks for the clarification.

I realize I have anecdotal evidence this is true, having charged my bicycle headlight off a laptop USB port many times.